Joint fingerprinting and decryption content security protection method for remote sensing images based on neighborhood similarity

Abstract

The joint fingerprinting and decryption (JFD) framework has high efficiency and can provide comprehensive, effective content security protection for remote sensing images. However, the fingerprint is generated by random partial decryption in JFD, which will lead to obvious degradation of image quality after embedding fingerprints and will severely impact the precision and further application of remote sensing images. In order to solve this problem, an improved JFD scheme based on neighborhood similarity is proposed in this paper, where the image is partitioned into different regions, and the regions that affect future application most are excluded from the fingerprint embedding area. Moreover, neighborhood similarity is defined in order to evaluate the change of pixel correlation, and areas with good neighborhood similarity after encryption are chosen to embed fingerprints. The experimental results prove that the fingerprinted image has a good image quality, and it will not affect the following application such as edge extraction and unsupervised classification, etc. Therefore, it is a suitable content security protection method for remote sensing image.